A forecast, written the year he left OpenAI, arguing that almost everyone was underrating both the speed and the physical scale of what was coming.
The call was good. The constraint on AI isn't models — it's power, memory and concrete. So he bought the bottleneck and shorted the software that agents were going to eat.
Shorting is the reverse trade: you borrow shares, sell them now, and buy them back later. You make money when the price falls — and you lose, without any limit, when it rises. Remember that. It is how the second half of this goes wrong.
439% net. A 165-page essay raised him $100m that November, from big names in tech. Billions followed. He was 22.
Other funds copied the trades. So much of the run in neoclouds and semis was his own crowd piling in behind him — wonderful, right until everyone needs the same door.
None of that saved him. Hold onto that.
Leverage is borrowing so you can bet more than you own. Four turns means four dollars of position for every dollar that's actually yours.
$10bn came in and became $30bn of equity — his money, and the only money that absorbs a loss. At four turns, that implies $120bn of positions — $90bn of it borrowed. Hold that number loosely.
Up, a 5% move lands as +20% on him. Down, a 25% move and he isn't hurt. He's gone.
A margin call isn't the market punishing you. It's your lender deciding your cushion got thin, and demanding cash you haven't got.
13F filings are public. Six names carried 85% of his disclosed book. He owned 8.2% of Core Scientific outright.
Read it as compliance and it's paperwork. Read it as a shopping list and it's something else. You cannot leave a position that size quietly.
Both legs lost at once. What he owned sold off, and what he'd shorted went up — which, on a short, is the direction that costs you. Two bets, both bleeding, same afternoon.
And once the street knows a man has to sell, it sells what he owns and shorts the rest. Nobody has to collude. Everybody can read.
On 24 July he wrote to investors. Up 439%. The sell-off, he said, had thrown up the most attractive opportunities since early 2025. Fresh capital welcome from 1 August.
The book was liquidated on 30 July.
Six days. What he knew when he wrote it isn't on the record and I'm not going to pretend otherwise. The dates are on the record.
As it rang round for help, the firm described what it was selling as hedges — modest positions held to limit risk. Buyers opened the file and found directional bets at scale, many of them against Adobe. Some passed on the spot.
Citadel took the book at a rumoured 20–50% discount. That discount is the entire trade.
They're market-neutral — already hedged against what killed him. Nobody holds $120bn forever, but they can bleed it out slowly. He had until Thursday. One had time, the other had a deadline. The price is the gap between those two facts.
The episode likens Griffin to Buffett in 2008, to JP Morgan in 1907 — the man who absorbs the risk before it becomes everyone's problem. Fair. Also worth saying plainly: the man who steadies the system ends up owning it cheap. Nobody rescued anybody. They priced it.
Note what didn't go. The banks took the public book because that's the part they could sell. The Anthropic stake — around $5bn, from a May round at $965bn, IPO possibly October — wasn't in the sale.
You cannot margin-call a private mark. Of the ≈$8bn left — down from about $30bn at the start of July — most is private stakes the firm values itself, and which the Times notes may or may not be worth what it hopes. The liquid money is gone. What survives is the part he gets to price.
One question decides it: was the failing thing collateral for everyone else?
LTCM was — the currency market couldn't absorb the exit. Bear Stearns was — its paper sat in everyone's trades. Archegos wasn't, and it still took Credit Suisse down eventually.
This one isn't. A stock drawdown is not a credit event. Not the same animal.
Though the comparison isn't mine. As the news broke, Daniel Loeb posted a link to When Genius Failed — the book about LTCM. The street reached for it inside an hour.
CORRECTION — the episode dates LTCM to 1994 and has it causing the Asian financial crisis. 1994 is when LTCM was founded; it collapsed in September 1998, and the Asian crisis ran through 1997, ahead of it. The argument still holds. The chart uses the corrected dates.
With the forced seller gone, there's no mechanical reason left to short any of it. Nasdaq +2.8% Thursday. Korea ripped. Wall Street is reading that as a reprieve rather than proof the thing is contained.
Everyone wants this to be the market calling time on AI. It isn't. It's a leverage story that happened to be wearing an AI-shaped book.
Four turns. Six names carrying the risk. Stakes too big to exit quietly. A short leg correlated to the long leg. Hedges built for the wrong disaster. A public filing telling everyone where to aim.
Those are construction failures. They'd have killed him just as fast betting on railways.
Though an investor document had promised no limits on what he'd buy, how concentrated it could get, or the amount of leverage he might use. Nobody was misled about the machine. They wired the money anyway.
Nothing in the reporting alleges fraud. Being catastrophically wrong with borrowed money is not a crime. It is a Tuesday.
The letter invited capital from 1 August. The book was sold on 30 July. The window opened after the fund was already gone.
His outcome and his investors' outcome were never the same outcome. Management fees came off the top the whole way up.
He raises again. Mid-twenties, no misconduct alleged, and 439% is permanently on his record — the tape doesn't care it was followed by a zero. The essay hasn't been refuted either; the buildout kept validating it while his book was being sold. The pitch writes itself: right about the thesis, wrong about the sizing, and I've had the education.
The precedent is exact. John Meriwether ran LTCM into a $3.6bn rescue by fourteen banks in 1998. He opened JWM Partners in 1999 with $250m — same strategy, less leverage. It lasted until 2009.
He has told investors he has stopped using borrowed money. He did not promise not to start again.
And at least one of them is still in. John Pfeffer, an investor since inception, went on the record: the leverage was overly aggressive, possibly existentially threatening — and he is absolutely delighted they are still in business.
Which is the part that should worry you. The asymmetry isn't a scandal. It's the product, and there is a queue.
Being right about the technology and staying solvent long enough to collect are two different jobs. He was exceptional at the first one.
Griffin was never smarter about AI than Aschenbrenner. He just never had to sell. That isn't a footnote — it's the entire skill, and it's the one nobody writes a 165-page essay about.
| Figure | Value | Source | Status |
|---|